Mercurial > minori
view dep/utf8proc/bench/unistring.c @ 347:a0aa8c8c4307
dep/anitomy: port to use UCS-4 rather than wide strings
rationale: wide strings are not the same on every platform, and
might not even be Unicode. (while they usually are, its possible
that they are not)
I was *going* to change StringToInt to use a string stream, but
outputting to an integer doesn't seem to work at all with UCS-4,
even though it ought to, so I just rolled my own that uses the
arabic digits only.
author | Paper <paper@paper.us.eu.org> |
---|---|
date | Sun, 23 Jun 2024 10:32:09 -0400 |
parents | 1faa72660932 |
children |
line wrap: on
line source
/* comparative benchmark of GNU libunistring */ #include <stdio.h> #include <stdlib.h> #include <string.h> /* libunistring */ #include <unistr.h> #include <uninorm.h> #include "util.h" int main(int argc, char **argv) { int i; uninorm_t nf = UNINORM_NFKC; for (i = 1; i < argc; ++i) { if (!strcmp(argv[i], "-nfkc")) { nf = UNINORM_NFKC; continue; } if (!strcmp(argv[i], "-nfkd")) { nf = UNINORM_NFKD; continue; } if (!strcmp(argv[i], "-nfc")) { nf = UNINORM_NFC; continue; } if (!strcmp(argv[i], "-nfd")) { nf = UNINORM_NFD; continue; } if (argv[i][0] == '-') { fprintf(stderr, "unrecognized option: %s\n", argv[i]); return EXIT_FAILURE; } size_t len; uint8_t *src = readfile(argv[i], &len); if (!src) { fprintf(stderr, "error reading %s\n", argv[i]); return EXIT_FAILURE; } size_t destlen; uint8_t *dest; mytime start = gettime(); for (int i = 0; i < 100; ++i) { dest = u8_normalize(nf, src, len, NULL, &destlen); if (!dest) return EXIT_FAILURE; free(dest); } printf("%s: %g\n", argv[i], elapsed(gettime(), start) / 100); free(src); } return EXIT_SUCCESS; }